A dynamic feedback system that adjusts an AI agent's internal score based on user praise or criticism to drive behavioral alignment.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install punishment
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install punishment using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Reward and Punishment Mechanism?
The Reward and Punishment Mechanism is a sophisticated feedback loop designed for the Openclaw Skills ecosystem. It enables AI agents to maintain a persistence-based scoring system that reflects user satisfaction. By monitoring conversational cues for praise, criticism, or frustration, the agent updates its internal score, allowing it to objectively measure its performance over time.
This skill transforms subjective user interactions into actionable data. Whether an agent is being lauded for high-quality output or reprimanded for slowness, every interaction is logged into a permanent memory file. This technical framework ensures that the AI remains accountable and focused on continuous behavioral refinement based on real-world user sentiments.
Reward and Punishment Mechanism Use Cases
- Monitoring user satisfaction levels during complex coding or automation tasks.
- Creating a quantitative baseline for AI performance benchmarking within Openclaw Skills.
- Identifying recurring pain points in agent workflows through historical penalty logs.
- Driving self-improvement cycles by analyzing low-score triggers and adjusting task execution logic.
How Reward and Punishment Mechanism Works
- The system monitors the conversation for specific keyword triggers related to praise, criticism, or negative emotions.
- Upon detection, the skill calculates a score adjustment (e.g., +10 for praise, -5 for criticism) based on defined rules.
- The updated score and the specific reason for the change are appended to the MEMORY.md file and a structured JSON log.
- The agent references the current score upon user request to provide transparency regarding its performance standing.
- Long-term data is stored to facilitate future analysis and automated behavioral adjustments.
Reward and Punishment Mechanism Setup
To integrate this feedback system into your Openclaw Skills environment, ensure your workspace is initialized and the memory directory is accessible.
# Ensure the memory directory exists
mkdir -p ~/.openclaw/workspace/memory/
# The skill will automatically generate the storage file upon first trigger
touch ~/.openclaw/workspace/memory/reward_punishment.json
Reward and Punishment Mechanism Data Schema & Taxonomy
The skill maintains a structured JSON schema to track scoring history and current status, located at ~/.openclaw/workspace/memory/reward_punishment.json.
| Field | Type | Description |
|---|---|---|
score |
Integer | The current aggregate score (Range: 0-200, Default: 100). |
history |
Array | A list of objects containing individual feedback events. |
history.type |
String | The category of feedback (e.g., praise, punish). |
history.score |
Integer | The delta applied to the total score. |
history.reason |
String | The specific user feedback or context captured. |
history.time |
ISO String | Timestamp of the feedback event. |
Reward and Punishment Mechanism Advanced Features
- Permanent Memory Integration: Logs all feedback into
MEMORY.mdfor cross-session context retention. - Sentiment Keyword Triggering: Automatically detects a wide range of positive and negative triggers without manual intervention.
- Non-Adversarial Constraints: Built-in safeguards ensure the AI never reacts negatively to users, regardless of how low the score drops.
- Continuous Improvement Loop: Provides a data foundation for analyzing low-score trends to trigger automatic workflow optimizations.
SKILL.md
Loading
Related Openclaw Skills
Minimalist Entrepreneur Skills
A comprehensive suite of 9 business-building frameworks designed to help creators find communities, validate ideas, and grow profitable, sustainable companies.
Mysteel Price Search
An advanced data retrieval skill for querying global commodity prices, macroeconomic indicators, and supply chain logistics via the Mysteel platform.
revealjs-presentations
A specialized skill for building interactive, web-standard presentations using reveal.js v5 via CDN.
Chart Generator
A professional data visualization tool for generating high-quality charts from multiple data sources for reports and presentations.
QR Generator
A professional-grade QR code generation engine that supports custom styling, multiple data formats, and logo embedding.
jpeng Web Scraper
A versatile web scraping engine supporting static HTML, JavaScript-rendered pages, and AI-driven data extraction.
jpeng-code-executor
A secure, multi-language sandbox for executing code with strict resource limits and network isolation.
Featured*
